Support ADT types in type info reflection #151142
Conversation
Support {bool,char,int,uint,float,str} primitive types for feature
`type_info` reflection.
SpriteOvO
added
A-type-system
rustbot
added
S-waiting-on-author
|
|
SpriteOvO
changed the title
SpriteOvO
force-pushed
the
type-info-adt
branch
from
1bd8449 to
0b20893
Compare
| sym::name => { | ||
| if let Some(name) = name { | ||
| let name_place = self.allocate_str_dedup(name.as_str())?; | ||
| let ptr = self.mplace_to_ref(&name_place)?; | ||
| self.write_immediate(*ptr, &field_place)? | ||
| } else { | ||
| let (variant, _) = self.downcast(&field_place, sym::None)?; | ||
| self.write_discriminant(variant, &field_place)?; | ||
| } | ||
| } |
There was a problem hiding this comment.
I feel like I should write sym::Some discriminant for name? 🤔 Writing ptr directly seems to work, but this relies on how Option<&'static str> is underlyingly represented IIUC.
But const eval interpreter doesn't complain about this -- perhaps it should?
There was a problem hiding this comment.
the interpreter doesn't care if the value is exactly the same as if it had its discriminant written. There is some talk about SetDiscriminant needing to happen always, but it's not really important for us here, this will be fixed by whoever changes it
also irrelevant if we make the field just always have a name (rendering the index)
|
The job Click to see the possible cause of the failure (guessed by this bot) |
| #[non_exhaustive] | ||
| #[unstable(feature = "type_info", issue = "146922")] | ||
| pub struct Field { | ||
| /// The name of the struct. |
There was a problem hiding this comment.
could also make it non-optional and produce integer names for tuples and tuple structs
There was a problem hiding this comment.
Oh that was a typo.
could also make it non-optional and produce integer names for tuples and tuple structs
Good idea.
| sym::name => { | ||
| if let Some(name) = name { | ||
| let name_place = self.allocate_str_dedup(name.as_str())?; | ||
| let ptr = self.mplace_to_ref(&name_place)?; | ||
| self.write_immediate(*ptr, &field_place)? | ||
| } else { | ||
| let (variant, _) = self.downcast(&field_place, sym::None)?; | ||
| self.write_discriminant(variant, &field_place)?; | ||
| } | ||
| } |
There was a problem hiding this comment.
the interpreter doesn't care if the value is exactly the same as if it had its discriminant written. There is some talk about SetDiscriminant needing to happen always, but it's not really important for us here, this will be fixed by whoever changes it
also irrelevant if we make the field just always have a name (rendering the index)
|
☔ The latest upstream changes (presumably #151210) made this pull request unmergeable. Please resolve the merge conflicts. |
4 participants
Tracking issue: #146922
#![feature(type_info)]This PR supports ADT types for feature
type_inforeflection.(It's still a draft PR, with implementation in progress)
Note that this PR does not take SemVer into consideration (I left a FIXME comment). As discussed earlier (comment), this requires further discussion. However, I hope we could get an initial implementation to land first, so we can start playing with it.
Progress / Checklist
mainbranch(It's currently based on PR Support primitives in type info reflection #151123, so here's an extra commit)
r? @oli-obk